GradientDescent

Mind the gap.

Learn optimization from first principles, then make it work at scale, on messy data, against the clock. Gradient Descent teaches both.

In class

  • Ten cities
  • Clean data
  • No time limit
  • One right answer

In production

  • Thousands of orders
  • Messy constraints
  • A hard deadline
  • Judged on cost
Most engineers meet production problems for the first time at work.
Coming soon

Two ways to learn

Guides are on the way. Until then, the best way to learn is to compete.

Explore competitions

Optimization from first principles

How to model a problem, why the algorithms work, and when to reach for each one. Built up from the basics.

Production-grade solvers

Code that runs fast, handles messy data, and holds up under real limits and deadlines.